Strategic Planning for Newspaper Delivery Problem Using Vehicle Routing Algorithm with Time Window (VRPTW)

brrrclergymanNetworking and Communications

Jul 18, 2012 (5 years and 11 months ago)


stract— We present a case study on the application of
techniques for solving the Vehicle Routing Problem with Time
Window (VRPTW) to improve the newspaper distribution
service in Bangkok, Thailand. The improvement of the
distribution activity is part of the strategic planning of the
company which aims to reduce number of staffs and to reduce
cost occurred in the distribution process. Empirical results
indicate that the heuristic method with some modification of
steps and constraints can find relatively good relations. It trends
to save cost of distribution up to 5.3% which bring service cost
from 23 % to 17.70%.

Index Terms— Vehicle Routing Problem, Sweep Algorithm,
Strategic Planning, Newspaper Distribution Problem (NDP).


The Vehicle Routing Problem was first introduced in
1959 by Dantzig and Ramser [1]. Their paper appeared in
the journal of Management Science concerning a fleet of
gasoline delivery trucks between terminal and a truck
number of service stations supplied by the terminal [2]. The
problem formulated in the Dantzig and Ramser’s paper
given the name “Dispatching Problem” and many years later
was coined the name “Dantzig and Ramser’s Problem” and
“Vehicle Routing Problem” respectively.
Ever since, the Vehicle Routing Problem (VRP) and its
variants have been intensively studied and received
considerable attention for many decades because of the
importance of mobility and its interest in different
application in logistics and supply chain management.
VRP is an optimization problem basically consists of finding
the set of routes with overall minimum total cost or total
travel times. The followings are some of its variant
constraint: every customer must be visited at once by

Manuscript received Feb 02, 2010. This work was supported and partially
inancially by University of the Thai Chamber of Commerce (UTCC),
Bangkok, 10400, Thailand.
A. Boonkleaw is now a second year student, Ph.D. Program in Logistics,
School of Engineering, University of the Thai Chamber of Commerce,
Bangkok, 10400, Thailand., (phone:+66818495088, fax:+6626923014,
e-mail:, ).
N. Suthikarnnarunai, PhD., is now Associate Dean, Graduate School
and Director, Ph.D. program in Logistics, School of Engineering,
University of the Thai Chamber of Commerce, Bangkok,10400, Thailand,
(phone:+66859071889, fax:+6626923014, email: ).
R. Srinon, PhD., is now Department Chair, Department of Logistics
Engineering, University of the Thai Chamber of Commerce, Bangkok,
10400, Thailand, (phone:+66842108109, fax:+6626923014, email: ).
vehicle, all demands must be served, overalls demands must
ot exceed the vehicle’s capacity, customer demands are
known, travel time is accurate.
VRP is an integer programming and is one of the
representative combinatorial optimizing problem and known
to be NP-hard and therefore difficult to solve [3]. The
traveling salesman problem (TSP) can be viewed as a
special case of VRP in which number of vehicle is one. TSP
has been intensively studied as well.
In this paper, the study of a newspaper distribution was
presented. The performance of initial routes was based on
data from a morning newspaper company located in
Bangkok, Thailand. We considered only the problem
concerning the distribution of newspaper between a
Distribution Center (DC) to drop points, not to final
Because getting the best results by applying the exact
algorithm from instance computer program is very costly,
therefore approximate solutions with sufficient accuracy are
often desired by the small and medium-sized enterprises.
The vehicle routing problem with time window (VRPTW)
was considered in this paper since customers often set
windows in which the vehicle has to arrive before they left
their home for work. For that, guaranteed on time delivery
and service quality are the main key success factors of
newspaper industry, which give the company competitive
This research was to develop VRP to serve morning
newspaper company’s strategic planning of minimizing total
traveling time and cost while increasing customer
satisfaction. The improvement is achieved by adjusting new
routs, reducing number of vehicles with result in lower total
In the last section, the routes performances achieved by
vehicle routing algorithm were then compared to that of initial
routes. Sensitivity analysis was also presented for alternative
solutions. Delivery time allowed and vehicle capacity were
the main criterion.

n this section, we briefly review the literature in general
of VRP with some of its variants - Capacitated Vehicle
uting Problem (CVRP), Vehicle Routing Problem with
Time Window (VRPTW), Multiple Depot Vehicle Routing
Problem (MDVRP), Vehicle Routing Problem with Pick-Up
and Delivery (VRPPD), Vehicle Routing Problem with
Strategic Planning for Newspaper Delivery
roblem Using Vehicle Routing Algorithm with
Time Window (VRPTW)
Arunya Boonkleaw, Nanthi Suthikarnnarunai, and Rawinkhan Srinon
Engineering Letters, 18:2, EL_18_2_09
(Advance online publication: 13 May 2010)


Backhauls (VRPB) - and an application of VRP in
newspaper industries.
A. Vehicle Routing Problem
he most general version of the VRP is the Capacitated
Vehicle Routing Problem (CVRP) which is a problem in
which all customers must be satisfied, all demands are
known, and all vehicles have identical, limited capacity and
are based at a central depot. The objectives are to minimize
the vehicle fleet and the sum of travel time while the total
demand of commodities for each route may not exceed the
capacity of the vehicle which serves that route [4],[5].
One of the most important extensions of the CVRP is the
Vehicle Routing Problem with Time Window ( VRPTW)
which is each customer must be served within a specific time
window. The objective is to minimize the vehicle fleet with
the sum of travel time and waiting time needed to supply all
customers in their required hour [4][6]. A variety of
algorithms including exact methods and efficient heuristics
have already been proposed for VRPTW.
Many researches have been studied intensively on heuristic
and meta-heuristic methods from 1995-2009 as summarized
in Table 1.

Table 1: Representative algorithms for Vehicle Routing
Problem with Time Window
Authors Year Methodologies
Dumas et all 1995 Time constraint routing and
heduling [7]
Liu and Shen

1999 Route-neighborhood-based
metaheuristic [8]
Bent et al. 2003 Two-stage hybrid algorithm [9]
Kima, et al. 2005 Capacitated clustering [10]
ysgaard 2006 Precedence constraints [11]
Russell and Chiang 2006 Robust solution methods [12]
Chena, Hsueh and
2009 An elaborated solution
Algorithms [13]
Li, et al. 2009 Lagrangian heuristic [14]

Multiple Depot Vehicle Routing Problem (MDVRP) is a
problem that customers can be served from several depots. If
the customers are clustered around depots, then the
distribution problem should be modeled as a set of
independent VRP. The objectives are to minimize the
vehicle fleet and the sum of travel time while the total
demand of commodities must be served from several depots.
Vehicle Routing Problem with Pick-Up and Delivery
(VRPPD) is a VRP with the possibility that customers’
returning of some commodities is contemplated. The
objectives are to minimize the vehicle fleet and the sum of
travel time, with the restriction that the vehicle must have
enough capacity for transporting the commodities to be
delivered and for picking up at customers to return them to
the depot.
Vehicle Routing Problem with Backhauls (VRPB) is a
VRP in which customers can demand or return some
commodities. The objective is to find such a set of routes
that minimized the total distances traveled.
In addition, Table 2 represents the various methods
applying in Exact Algorithm, Classical Heuristic Algorithms
and Metaheuristic Algorithms.
Table 2: Representative Exact algorithms, Classical Heuristic
Algorithms and Metaheuristic Algorithms for VRP

a) Exact Algorithms
Authors Year Methodologies
Christofides and Eilon 1969 Branch and Bound [15]
Miller 1995 Branch and Bound [16]
Hadjiconstantinou et al 1995 Set-partitioning [17]
Baldacci et al. 2004 Branch and Cut [18]

b) Classical Heuristic Algorithms
Authors Year Methodologies
Dantzig and Ramser

1959 Constructive algorithm. First
approach [1]
Clarke and Wright 1964 Saving. Concurrent & sequential
Wren and Holliday

1972 Sweep Algorithm. Multiple depots
Lin and Kernighan

1973 Single-route improvement
Sequential k-exchange [21]
Gillett and Miller 1974 Sweep Algorithm. Single depot [22]
Foster and Ryan

1976 Petal algorithm. Optimal petal
solution [23]
Mole and Jameson

1976 Sequential Route-Building
Insertion position check [24]
Christofides et al.

1979 Sequential Route-Building
Sequential & Parallel construction
Fisher and Jaikumar

1981 Cluster-First Route-Second
Generalized Assignment + TSP [26]
Beasley 1983 Route-First Cluster-Second [27]
Altinkemer and
1991 Matching Algorithm. Matching
clusters [28]
Ryan et al. 1993 Petal algorithm [29]
Thompson and
1993 Multiple-Route Improvement
b-cyclic k-transfer [30]
Potvin and Rousseau

1995 Single-route improvement. Based
on 2-opt [31]
Bramel and
1995 Cluster-First Route-Second [32]
Renaud et al. 1996 Single-route improvement [33]
Kindervater and
1997 Multiple-Route Improvement [34]

c) Metaheuristic Algorithms
Authors Year Methodologies
Osman 1993 SA [35]
Taillard 1993 TS [36]
Gendreau et al. 1994 TS [37]
Van Breedam 1995 SA[38]
Rochat and Taillard 1995 TS [39]
Xu and Kelly 1996 TS [40]
Kawamura et al. 1998 ACO [41]
Bullnheimer et al. 1999 ACO [42]
Toth and Vigo 2003 TS [43]
Baker and Ayechew 2003 GA [44]
Mazzeo and Loiseau 2004 ACO [45]

B. Vehicle Routing Problem in Newspaper Industries
any researches on this area indicated that VRP was and
is still a great tool for minimizing the total cost of delivery
or total travel time in the newspaper industry. Table 3
presents some of heuristic VRP specifically to newspaper
industries during 1964-2008.
Engineering Letters, 18:2, EL_18_2_09
(Advance online publication: 13 May 2010)


Table 3: Representative Algorithms for the Vehicle Routing
Problem (VRP) in Newspaper industry from 1964-2008
uthors Years Methodologies
Clarke, J.W. Wright 1964 Savings technique [19]
Floyd 1967 Algorithm 97 shortest
Daganzo 1981 The Distance Traveled to
Visit N-point with C-stop[47]
Ree and Yun 1996 Branch and Bound [48]
Hurter, M. Van Buer 1996 Greedy and Or-Opt route
improvement heuristic[49]
Van Buer, Woodruff,
and Olson
1999 Metaheuristics-simulated
annealing /tabu search [50]
Carter and Ragsdale 2002 GA [51]
Song, Lee &. Kim 2002 Regret Distance Calculation
algorithm [52]
Russell, et al. 2008 Open vehicle routing problem
and zoning constraints [53]

There were many researches studied about production and
distribution process, but not many about preprint process.
Several literatures are summarized as below:

- Preprint Process
A genetic algorithm (GA) was used to approach to the
pre-print advertising scheduling problem and computational
results using data from a mid-size newspaper. The result
showed that the GA approach to developing schedules
reduces the processing time associated with creating the
preprint packages [51].

- Production and Distribution Process
A newspaper distribution problem for a metropolitan daily
Korean newspaper was also studied and then developed a
livery plan using a branch-and-bound heuristic with
simulated annealing (SA) [48].
Before that [49] develop a deterministic approach to a
dium sized newspaper production/distribution problem in
which they employ a greedy heuristic followed by an Or-Opt
route improvement heuristic. The problem was smaller and
involved only one printing press and more importantly
considered only a single product delivery to each zone.
Thus, each zone contained its own routing problem.
Also, Regret Distance Calculation algorithm was selected
for agent allocation, a Modified Urgent Route First
algorithm for vehicle scheduling, and a Weighted Savings
algorithm for routing in addressing the optimal agent
allocation, vehicle scheduling and routing for a major
newspaper in Korea, the experiment showed that the
formulation could significantly reduce delivery costs and
delays [52].
Other examples included [53] work on the open VRP,
their solution to the open vehicle routing problem and
zoning constraints (OVRPTWZC) showed significant
improvement in both the number of vehicles employed and
the total distance traveled over the existing operations of a
U.S. metropolitan newspaper.
A Dutch regional newspaper’s distribution process was
also studied [54] and the process was modeled by
constructing the travel time matrix using [46] algorithm and
[19] savings technique as the vehicle routing heuristic. In
[47], a newspaper delivery problem for the city of San
Francisco was considered as an application of a formulation
developed for predicting the distance traveled by fleets of
vehicles in distribution problems.
The formulation was a variant of the “cluster-first,
route-second” approach to solve vehicle routing problems.
In a follow up to [47] work, [50] extended the solution
method to include metaheuristics, simulated annealing and
tabu search. Its approach was deterministic and one of the
main findings was that recycling trucks to create more routes
while using fewer vehicles can lead to significant cost
Moreover, references [55], [56] applied VRP to school
bus routing. Other applications included inventory and
vehicle routing in the dairy food [57], transportation service
at university [58], public library system [59], post service
[60], and grocery delivery [61], waste collection [62],
magazine [63].

n real world, fleet of transportation is very complicated.
Number of trips, links/path and cost are to be considered.
Transportation often involves routing vehicles according to
customer given time allows that determines the customer’s
satisfaction level. Therefore, all publishers intensively
improve and adjust company’ strategies by pertaining their
internal resources with external resource (the market). The
competitive advantage can be achieved by concentrating all
the available resources on one basic strategy which is shorten
delivery time. The short delivery time if administered
efficiently and effectively could also result in less distribution
cost. This may be the ultimate choice since a declining
enterprise had difficulty to increase sales [64], [65].

a) Newspaper Industry Situation
Newspapers in Thailand and others countries are
confronting the most serious situation of their history. They
are facing an unprecedented famine in news print. Because
of declining circulation, many publishers are now carrying
printing and distribution cost per copy more than they
formerly did.
In order to keep all expenses and costs as budgeted, many
owners and investors have decided a general increase in
advertising rate and subscription rates. The reasons are to
keeps the smaller newspaper from being forced to suspend
and be able to compete with other off-line services and with
other media such as TV, radio and other on line services.
Moreover, publisher needs to improve the production and
distribution process as well as other processes in the
Currently, end consumer or reader prefers to consume
news from websites like Google or Yahoo and other search
engines. Those sources are sharing revenue and customers in
the market. They are also mimicking the job of editors by
using sophisticated computer programs to automatically
compile links to content from newspapers, wire services,
blogs, and other sources from around the world.
The newspaper companies cannot print the news sections
Engineering Letters, 18:2, EL_18_2_09
(Advance online publication: 13 May 2010)


of the newspaper in advance because of the requirement that
news be timely. Achieving on time delivery therefore is quite
challenging because publisher wants to delay their printing
as much as possible in order to get the latest news into the
prints. This gives delivery department time window as little
as three hours or less to get the papers to the readers.
Consequently, the logistics staffs have to work as fast as
possible. They also have to provide more available trucks
than needed and that cost the company more than it should.
Traditionally, the idea of low cost and fast delivery could
not be attained, simultaneously. It is a belief that some kind
of trade-off is necessary; the more of one advantage means
less of another. However, it was suggested that “seeking
reduction, either time or cost reduction is often the rewards”

b) Newspaper Category
Based on customer demand, newspaper can be categorized
into two parts, subscribing newspaper and non- subscribing
For non-subscribing newspaper delivery, demand is
probabilistic, so it needs an estimation to determine the
amount of newspaper to be printed and distributed. Waste of
unsold newspapers would happen if the amount of distributing
newspapers were bigger than sold newspapers and that also
cause the problem of reverse logistics. Therefore, the
distribution problem is not only about a delivery problem but
also about collecting back unsold copies on a daily or weekly
basis. In effect, about 50% of the distributed copies were
returned to the warehouse. Because of internal audit policy,
warehouse must provide enough space to keep unsold copies
until all are checked by internal auditor. Also, delivery
document must be promptly signed by outlet staffs authorized
by shop owners. Without signed document, cost from
complaining may occur if there is loss of whole pack of paper.
Consequently, cost of making a delivery of replacement to
those outlets would occur.
For Subscribing Newspapers Delivery, Demand of
subscribing newspaper is deterministic; it does not need to be
estimated for distribution. However, time allowed is a very
crucial issue in this type of delivery.

c) Newspaper Distribution Models and Lead Time
There are 4 models for newspaper distribution system in
Thailand as shown in Figure 1 and described below.
Model 1: In case that loading dock and DC are not located
in the same place, Newspapers will be transferred to DC for
packing, then delivered to drop points and then carrier will
distribute to final customers.
Model 2: In case that loading dock and DC are located in
the same place with packing required, newspaper will be
packed immediately, then delivered to drop points and then
carrier will distribute to final customers.
Model 3: In case that loading dock and DC are located in
the same place without packing required, newspaper will be
directly loaded into vehicle and delivered to drop points, and
then carrier will distribute to final customers.
Mode 4: In case that the final customers’ residents are
close to the printing house, carrier will pick newspaper at
loading dock and deliver to designated customers.
Loading Dock O O O O
ransportation O
Distribution C

Transportation O O O
Carrier O O O O
Reader O O O O

Figure 1: Models of newspaper distribution system in

Figure 2 shows current processes and lead time of physical
newspaper distribution. The processes involve loading,
transport, and carrier. How early the distribution process
could start depending on when the printing starts and how
efficient the mail room activities are. The printing and mail
room processes (or production process) will not be deliberate
in this paper. However, all processes have to be planned to
meet the delivery deadline. When the delivery deadline is set,
say at 6.30 am, an analysis of the process time for each
process will generate the required starting times. The key
timing parameters are the start of the press run, say at 1.00 am.
The carriers have to finish by the delivery deadline. The
transports have to depart from the loading dock at planed
departure times in order to arrive at the drop-offs before the
carriers start to deliver to their first customer or reader.

Figure 2: Processes and lead time to be considered for
physical newspaper distribution

d) Newspaper Delivery Problem
The Newspaper Distribution Problem (NDP) involves the
downstream movement of newspaper from the printing
process to the hand of readers. The NDP can be viewed as a
hierarchical distribution problem. That means the newspaper
delivery involves at least two distinct stages. The first stage
is from the production facility to the transfer points and the
second stage is from the transfer points to customers [39].
NDP is an example of a perishable-good production and
distribution problem. People who are working in publishing
companies classify physical newspaper as perishable goods
because they could be lost in significant value if delivered
late or over printed [32].
NDP is also vital in the newspaper industry provided that
it is directly tied to customer service level. Late delivery of a
newspaper may result in the loss of customers or may result
in the shutting down of a production line if numbers of
customers are rapidly reduced [5].

Distribution Start
Delivery Deadline
3.30 am
6.30 am
Pack& Load
Carrier to customer
4.30 am
Engineering Letters, 18:2, EL_18_2_09
(Advance online publication: 13 May 2010)


During the global financial crisis in 2009, newspaper and
agazine businesses, particularly medium-sized companies
are in need of a revision of the current business strategy,
which will not only allow businesses to maintain their
long-term potentiality, but also facilitate future growth and
also need to Improve the efficiency of the business process
to make the company more attractive to potential customers
or subscribers.
For a newspaper studied in this paper, local advertisement
accounted for 80% of newspaper revenues. Subscriptions
and newsstand sales make up most of the rest. The
traditional advertising model for years generated healthy
profits, subsidized all expensive expenses and help keep
subscription and newsstand prices low as the results of
gaining number of readers. But the model is breaking down.
Newspapers have a dramatic drop in advertising during the
recession. Newspaper ads revenue fell by 20% in the first
and second quarter of 2009 compared to a year earlier.
As mentioned earlier, traditional printed newspapers take
a financial beating, new media sources are rising on the
Internet, helped by low entry costs. Alternative sources of
news and information are becoming available to a potentially
vast audience via an increasing number of wired and
wireless devices. So far, in the company, no one has come
up with a workable strategy for garnering sufficient
advertising or subscription revenues on the Internet. For
short term plan, however they intend to increase subscription
prices for their traditional print editions. The profit per order
will definitely be increased and the main reason is to keep
the bottom line in black ink.

a) General Strategies for newspaper’s company
The followings are the main strategies for a newspaper
company so employee and management have to pay much
attention to the strategies to get customer satisfaction.
Product: Experienced auditors will review and develop
the content of newspaper every year in order to response the
need of customers.
Price: For subscription sales, discount and premium are
always offered in order to increase their sale volumes. For
non-subscription sales, discount will be offered to agents or
bookshops, not direct to the reader. The price movement in
the domestic market is also monitored closely and a
competitive price is always set to compete with their

Table 4: Newspaper price model in 2009
(Unit: Baht*)
Condition Newsstand Subscription Save
Per issue 25 13.4 46%
1-year rate 9,125 4,900 46%
2-year rate 18,250 9,500 48%
* Exchange rate was 34 Baht / USD as of 1 Feb 2010

able 4 shows the comparison between newsstand and
subscription prices of a morning newspaper in Bangkok.
One-year subscription rate is offered about 46 % less than
newsstand price whereas two-year subscription rate is
offered with more discount at 48 %. Premium, free gift and
free reading copy are also additionally offered in certain of
period. Free home delivery is offered as a complimentary to
all readers.
Distribution Channel: Table 5 shows the proportion of
sales with allowed conditions. The Company sells
newspapers directly to their customers by attending Bangkok
International Book Fair and Book Expo Thailand which are
held in March and October yearly in Thailand. Meanwhile,
Bookshop is the main distribution channel for newspapers.
Internet is another mean for the company to present their
products via the company web site. Customers can purchase
the newspaper in bookshops, place their orders by using mail
order or they can order directly at company call center. For
delivery time, subscribers would like to have their copy
before leaving home to work while for bookshops, the
newspapers must be arrived before shops opening.

Table 5: Proportion of sales with allowed sales condition
in each channel.
Channel % Sales Sales Condition
Direct Sales 50% Subscription rate
Telesales 40 % Subscription rate
Bookshop 5% Subscription rate
Internet 4% Subscription rate
Book Fair 1% Newsstand rate

Process of Customer Complaints handling: Responding
to customer complaints is the major concern of the company.
Customers will receive the answer back within 1 hour after
the placed complaints and customer will receive replaced
paper within 3 hours if paper cannot be found. Figure 3
show the detail of customer complaint’s process.

Figure 3: Process of Customer complaint handling

1. Customer call CRM to complain non-received paper
2. CRM report logistics department via computer
3. Logistic assign staff and report back to CRM the
estimated arrival time.
4. CRM call to inform customer.
5. Newsboy delivers special newspaper to customer and
report back to Logistics Department when task is

1) complain for
2) report via
3) assign staff
4) arrange delivery
Information flow

Physical delivery

Engineering Letters, 18:2, EL_18_2_09
(Advance online publication: 13 May 2010)


Two main objectives that a newspaper company sets for
their annual strategic business plan are to achieve total
annual sales and profit, and to meet customer satisfactions.
To gain more profit, all cost has to be monitored and
controlled under consideration of customer satisfaction
fulfillment. The following section, distribution cost which
consists mainly of labor cost and also transportation cost will
be discussed in detail.

b) Newspaper Distribution Cost
The distribution of newspaper has a number of features
distinguishing it from other distribution operations because
newspaper is not to be produced in advance for inventory
keeping. As a result, distribution centers play an important
role, while print production and distribution are necessarily
intimately related. In addition, the total time devoted to both
production and to distribution may be severely limited,
thereby further tying together the design and operation of the
production and distribution functions. These distinguishing
features increase the complexity of the production and
distribution problem for newspaper [32].
Table 6 shows the strategic planning that company target
for future distribution. The newspaper company aims to
reduce distribution cost from 23% down to 20% and
complain rate from 0.05% to 0.45 % in following year.

Table 6: Strategic planning for a newspaper distribution
Strategic planning Current Target
Provide service at lower cost 23.0% 20.0%
Improve service quality by
reducing delivery complain
0.050% 0.045%

Newspaper distribution cost can be divided into two
major categories: first, costs associated with the actual
production and distribution activities and second, costs
directly attributable to the perish ability of either an input or
an output. In the newspaper problem, an infeasible solution
would be one in which not all newspapers are delivered by
the deadline. The costs associated with this infeasibility may
include the cost of the newspaper, a lost subscription, the
cost of processing the complaint, the cost of making a
special delivery, etc.

Table 7: Component of newspaper distribution cost.
(Unit: Baht*/ day)
Activity Pack & Load Transport Total
No. Staff 4 7 11
No. of Vehicle 7 7
1.Staff exp. 800 2,100 2,900
2.Distribution exp. 4,330 4,330
2.1 Petrol 3,500 3,500
2.2 Toll way 630 630
2.3 Maintenance & other 200 200
* Exchange rate was 34 Baht / USD as of 1 Feb 2010

Distribution cost in newspaper industry is a major expense
or the newspaper, making up to approximately 23 % of the
total cost. As shown in Table 6 and 7, distribution cost
consists of labor cost at 40% while transport and petrol costs
are account for 60% of distribution cost.
Labor Cost: In distribution department, labor cost is mainly
from packing staffs, loading staff, and drivers. In studied
company, there is no material handling equipment to support
in packing or loading process, therefore company requires a
number of labors to produce its goods and service. Labor
costs are considered variable which gives labor-intensive
industries as newspaper’s company an advantage in cutting
expenses during market downturns by controlling the size of
the employee base.
Beside distribution department, people from the finance,
marketing, operations, and human resources departments are
often self-directed. That is, they are given broad objectives,
but not specific directives. Decision making within a team is
usually based on consensus. Each department does not well
derstand the common objectives of the company. Up until
recently, decision making flowed in one direction so each of
department does not see the big picture of the company.
Therefore, Cross function team can be of help the company
improve employee performance and increase their multi
skills that is an initial way to save labor cost. Consensus
should be created within the team through an interactive
process. Ideally, when all department and management level
are working as a team, direction of increasing revenue is
well cleared and understood, all actions will be performed
by following company direction and definitely percentage of
expenses will be automatically decreased.
Transportation Cost: Transportation cost in studied
company includes petrol cost, toll way fee and maintenance
cost. Further, transportation costs have recently spiked due
to increases in gasoline and diesel fuel prices. To rise and
therefore delivery costs will trend higher. A method for
reducing newspaper distribution costs is by delivering more
products to the destinations during each delivery trip without
affecting constraints. In addition, multiple uses of vehicles
can also reduce transportation cost, yet implemented in
studied newspaper company. Currently, single use of
vehicles only for newspaper delivery is still a common
In addition, newspaper distribution cost analysis should be
done from time to time. Newspaper distribution cost analysis
is a technique which examines in detail all the costs incurred
in process of printing, loading, packing and carrying of
newspaper to the customer, it involves a study of cost control
which is directly applicable to the whole delivery operations.
In particular, by showing the degree of expense that each
part of the printing and delivery activity attracts, it helps to
improve the delivery policy of the company.
Newspaper’s company must provide just-in-time delivery
services to their customers. Typically those customers or
subscribers require daily. It is a significant management
challenge to design and develop an efficient delivery
schedule to meet demand. In other word, newspaper’s
company can take advantage of the just-in-time (JIT)
approach to achieve goals such as cost reduction, lead-time
reduction, quality assurance, and respect for humanity. Since
e performance of the publisher can be evaluated by various
criteria including lead times, on-time delivery, delivery
reliability, quality, and cost , deploying the JIT system is
crucial in improving customer satisfaction.
Engineering Letters, 18:2, EL_18_2_09
(Advance online publication: 13 May 2010)


There are many strategies planning studied report in
iterature. Reference [27] addressed route planning for
magazine and newspaper wholesalers. Other report include
cooperate strategy for express delivery services [40], and
JIT and TQM studied for improving delivery performance

elow are the assumption lists that shaped our formulation
of the VRP model.
1. Each route will start from and end at the Depot.
2. The cost of a route is proportional to the time
3. Travel times between each stop are known and
4. Demands (i.e., number of copies) at each of the
stops are known.
5. Unloading time per stop is constant for every stop.
6. The demand at each stop cannot split.
Constraints in this problem are
1. Total of 7 vehicles are available.
2. Hour of operations: there are time window of t
=180 minutes for delivering newspapers to the last
stops/customer and t=60 minutes for returning to
However, there are other constraints that did not define in
this paper due to intangible factor which cannot be part of the
model i.e. vehicle capacity which all copies shall be stored
behind the truck with door closing at all time of running. The
paper shall not be stored in the front seat or the roof of the

mathematical explanation of VRP for newspaper
delivery problem in this case may be defined as follows. Let G
= (V, A) be a network where V = {0, 1, …, n} is the vertex set
and A ⊆ V×V is the arc set. Vertex 0 is the depot and V\{0} is
the set of locations on the road network. Associated with
vertex i ∈ V\{0} is a non-negative demand d
. The parameter
represents a non-negative cost (traveling time in this case)
etween vertices i and j. The parameters K and U
are the
umber of vehicles and the capacity of vehicle k, respectively.
A three-index integer programming formulation will be
presented here where binary variables x
count the number of
imes arc (i,j) ∈ A is traversed by vehicle k (k = 1,…,K) in the
optimal solution. In addition, there are binary variables y
∈ V;
k = 1,…,K) that take a value of 1 if vertex i is visited by
vehicle k in the optimal solution and take a value of 0,
otherwise. The formulation is as follows:

∑ ∑∑
∈ ∈=
Aji Vi
),( 1
1.0 (1)

ubject to

}0{\1 (2)




KkViyx (5)

KkUyd (6)
∑ ∑∑
∈ ∈=
Aji Vi
),( 1
1.0 Kk,...,1

Si iSj

∈ ∈



,...,1,),(}1,0{ =∈∀∈ (10)

Equation (1) represents the objective function of this
problem to minimize total travel time of the operations.
Constraints (2) - (5) ensure that each customer is visited
exactly once, that K vehicles leave the depot, and that the
same vehicle enters and leaves a given customer vertex,
respectively. Constraints (6) are the capacity restrictions for
each vehicle k, whereas constraint (7) is a time window
constraint. The unloading time also presented here in (7).
The sub-tour elimination constraint for each vehicle is
shown in constraint (8).

he sweep algorithm is 2-phase algorithm [38]. The
problem is decomposed into its two natural components:
Clustering of vertices into feasible routes, then actual route
construction, in other word cluster first and route second
algorithm. The sweep algorithm applies to planar instances of
the VRP. It consists of two parts[5],[57],[58]:
1. Split: Feasible clusters are initialed formed rotating a ray
centered at the depot.
2. TSP: A vehicle routing is then obtained for each cluster
by solving a TSP.
The generic sweep algorithm uses the following steps
1. Locate the depot as the center of the two –
dimensional plane.
2. Determine all the polar coordinate of each stop
with respect to the depot.
3. Start sweeping all customers by increasing polar
Engineering Letters, 18:2, EL_18_2_09
(Advance online publication: 13 May 2010)


4. Assign each customer encompassed by the sweep
to the current cluster.
5. Stop the sweep when adding the next stop would
violate the maximum vehicle capacity.
6. Create a new cluster by resuming the sweep where
the last one left off.
7. Repeat steps 4-6, until all customers have been
included in a cluster.
The original sweep method, as mentioned above, has the
vehicle capacity and the travel time to next stop as the route
termination rules. Figure 4 illustrates a clustering process.

Figure 4: Clustering process: clockwise manner

In this research, the vehicle capacity constraint is still
hold. However, it allows the sweep to skip a stop when the
travel time to that stop would exceed the time limit. The next
stop after the skipped stop will be tested by the same
termination rules. If it exceeds the capacity then the sweep
terminates, and the stop which has the least angle which is
not include in any cluster yet will be used as the starting stop
in the next cluster. The sweep considers the stops in
increasing angle until one is found that does not violate the
time limit. If no such stop is found, the cluster is terminated
and the next cluster is started at the stop with lowest degree
angle which has not been included in previous cluster yet.

n this section we will present the initial results achieved
after the skipped sweep method has been applied and
compare with the previous one. This research comprises 118
nodes, with different demand values ranging from 22 to 235.
The vehicle capacity is 2,000. The solution reveals that
seven vehicles will be used with total delivery time of 1,205
minutes or 10.21 minute per node in average.
Table 8 summarizes the results achieved by applying the
Modified Sweep Method. Every customer remains received
newspaper as time promised even apart of the customers, 49
drop points or representing 41.5%, received newspaper later
initial routes.
After we monitored the initial results, we found that there
were many opportunities and possibilities to improve the
results for example, reallocating drop points, adjusting lead
time without violating capacity constraints.

Table 8: Comparison initial results of newspaper delivery
between initial routes and new routes.
Stop points Total Travel Time
( minute)
No. of copies
Initial New Initial New Initial New
1 20 22 176 179 1,265 1,800
2 15 20 145 166 1,311 1,742
3 17 19 165 172 1,153 1,140
4 17 15 156 177 1,372 1,115
5 19 20 168 176 1,451 1,597
6 19 18 153 173 1,593 1,621
7 11 4 142 162 1,211 341
Total 118 118 1,105 1,205 9,356 9,356

In addition, when analyzed the result in other aspects, we
ound that the 7th route has to be delivered only 4 drops and
consumed total time at 162 minutes with total capacity of
341 copies only. Therefore, we strongly believe that there
were still rooms for improving the results. Sensitivity
analysis will be proposed for reducing number of vehicle
which is part of minimizing total cost of using vehicle as
targeted in company’s strategic planning.
Considering company’s current operations, we determined
that the t value could be larger than the stated time window.
The heuristics results when we assigned the wider time
window revealed that at t equal to 200 minutes gave a
preferable result. Only 6 vehicles will be needed in the
delivery process. However, total travel time of every route
exceed the previous stated time window. Therefore, we
adopted traveling salesman problem method to order the
drop points in each route.
Table 9 shows the comparable result of total travel time
obtained by skipped sweep algorithm only and improving
results after using TSP method. It shows that only 2 routes
used time to operate in distribution activity more than
previous stated time window of 180 minutes.

Table 9: The results of total travel time (minute) obtained
by skipped sweep algorithm

No. of
Skipped Sweep
Algorithm with Time
Window 200 minutes
After Ordering
Using Traveling
Salesman Problem
1 25 198 179
2 20 192 173
3 16 188 180
4 19 196 177
5 12 199 189
6 26 196 188
Total 118 1,169 1,086

n this paper, the two aspects of research area were
studied, strategic planning and algorithm, for a morning
newspaper in Bangkok which aim to improve delivery
service within time allowed.
For the first aspect, strategic planning for newspaper’s
company was studied and discussed specifically to price
model, labor cost and distribution cost. Subscription and
newsstand prices should be set as low as possible when
advertising sales generated healthy revenue and profits.
Engineering Letters, 18:2, EL_18_2_09
(Advance online publication: 13 May 2010)


However, in the crisis situation when newspaper ads
dropped and affect the bottom line of the company,
increasing subscription prices should be taken into
consideration. Transportation cost can be improved by
reducing traveling lead time and by multiple uses of vehicle.
Transportation cost should be monitored continuously.
For the second aspect, the experiment has been developed
twice for vehicle routing algorithm. The first result showed
that developing a vehicle routing algorithm to solve variant
VRPTW remains unsolved as modified sweep algorithm
resulted in the delay of delivery time for 41.5% of all drop
points. However, to assure readers satisfaction on every drop
point, a good distribution and precise amount distribution
remains meet on time delivery.
Moreover, sensitivity analysis was conducted as to improve
results by increasing time allowed from 180 minutes to 200
minutes in sweep algorithm steps. The results indicate that the
heuristic method with some modification of steps and
constraints can find relatively good relations. It trends to save
cost of distribution up to 5.3% which bring service cost from
23 % to 17.70%.
The following study areas will be the direction of future
1. Scheduling of production process.
2. Relocating drop points
3. In this paper, only DC to drop points was solved,
therefore to complete all process of distribution
network, drop points to end customers should be
considered as multi-time window, multi- customer

he author thanks anonymous reviewers for their support
and valuable comments which led to an improved paper. A
preliminary version of the research appeared in [5].

[1] G.B. Dantizig, J.H. Ramser, ”The Truck Dispatching Problem”
Management Science, 1959, Vol. 6, pp 81-91.
[2] B.Kalleharge, “Formulation and Exact Algorithms for The Vehicle
Routing Problem With Time Window” Computers and Operation
Reserrch 35, 2008, pp.2307-2330.
[3] H. Hashimoto et al, “ An Iterate Local Search Algoritm for the
Time-Dependent Vehicle Routing Problem with Time Windows”
Discrete Optimization 5, 2008, pp. 434-456.
[4] Avalibale:
ons/VRPPDDesc.html 10/18/2008.
[5] A. Boonkleaw, S. Suthikannarunai, R. Srinon, “ Strategic Planning
and Vehicle Routing Algorithm for Newspaper Delivery Problem:
Case Study of Morning Newsppaper, Bangkok, Thailand, Proceeding
of the World Congress on Engineering and Computer Science,
Sanfranciso, USA, Vol. 2, 2009.
[6] P. Mochado, et all, “Vehicle Routing Problem: Doing the
Evolutionary Way”
[7] Desrosiers, J., Y. Dumas, M.M. Solomon and F. Soumis (1995),
“Time Constrained Routing and Scheduling”, In Handbooks in
Operations Research and Management Science, Vol. 8. Network
Routing, M.O. Ball, T.L. Magnanti, C.L. Monma, G.L. Nemhauser
(eds), North-Holland, Amsterdam, pp. 35−139.
[8] Fuh-Hwa Franklin Liu *, Sheng-Yuan Shen “A
route-Neighborhood-Based Metaheuristic for Vehicle Routing
Problem With Time Windows, European Journal of Operational
Research 118 (1999) 485-504.
[9] R. Bent, P.V. Hentenryck, “A Two-Stage Hybrid Algorithm For
ckup and Delivery Vehicle Routing Problems With Time
Windows”, Computers & Operations Research 33 (2006) 875–893.
0] B. Kima., S. Kimb, S. Sahoob, “Waste Collection Vehicle Routing
Problem With Time wWndows, Computers & Operations Research
(2006), 3624–3642.
1] Jens Lysgaard, “Reachability Cuts for the Vehicle Routing Problem
with Time Windows”, European Journal of Operational Research
175 (2006) 210–223.
[12] Robert A. Russell, Wen-Chyuan Chiang, “Scatter Search For The
Vehicle Routing Problem With Time Windows” European Journal of
Operational Research 169 (2006) 606–622..
[13] Huey-Kuo Chen, Che-Fu Hsueh,., Mei-Shiang Chang, “Production
Scheduling and Vehicle Routing with Time Windows for Perishable
Food Products”, Computers & Operations Research 36 (2009) 2311
– 2319.
[14] Jing-Quan Li a,*, Pitu B. Mirchandanib, Denis Borenstein, Real-time
Vehicle Rerouting Problems With Time Windows, European Journal
of Operational Research 194 (2009) 711–727.
[15] Christofides, N. and S. Eilon (1969, “September). An algorithm for
the vehicle-dispatching problem”. Operational Research Quarterly
20(3), 1969, pp.309–318.
[16] Miller, D. L., “A matching based exact algorithm for capacities
vehicle routing problems”. ORSA Journal on Computing 7(1), 1995,
[17] Hadjiconstantinou, E., N. Christofides, and A. Mingozzi, “A new
exact algorithm for the vehicle routing problem based on q-paths and
k-shortest paths relaxation”. Annals of Operations Research 61(1-4),
1995, pp.21–43.
[18] Baldacci, R., E. Hadjiconstantinou, and A. Mingozzi, “An exact
algorithm for the capacitated vehicle routing problem based on a
two-commodity network flow formulation”..Operations Research
52(5),2004, pp.723–738.
[19] Clarke, G. and J. Wright, “Scheduling of vehicles from a central
depot to a number of delivery points”. Operations Research
12(4),1964, pp. 568–581.
[20] Wren, A. and A. Holliday, “Computer scheduling of vehicles from
one or more depots to a number of delivery points”. Operational
Research Quarterly 23(3), 1972, pp.333–344.
[21] Lin, S. and B. W. Kernighan, “An effective heuristic algorithm for
the traveling salesman problem”. Operations Research 21(2), 1973,
[22] Gillett, B. and L. Miller, “A heuristic algorithm for the
vehicle-dispatch problem”.Operations Research 22(2),1974, pp.
[23] Foster, B. A. and D. M. Ryan, “An integer programming approach to
the vehicle scheduling problem”. Operational Research Quarterly
27(2), 1976, pp. 367–384.
[24] Mole, R. H. and S. R. Jameson, “A sequential route-building
algorithm employing a generalized savings criterion”. Operational
Research Quarterly 27(2), 1976, pp.03–511.
[25] Christofides, N., A. Mingozzi, and P. Toth, “The vehicle routing
problem”. In N. Christofides, A. Mingozzi, P. Toth, and C. Sandi
(Eds.), Combinatorial optimization, Chapter 11, 1979, pp. 315–338.
Chichester, England: John Wiley & Sons Ltd.
[26] Fisher, M. and R. Jaikumar, “A generalized assignment heuristic for
vehicle routing”. Networks11(2),1981, pp.109–124.
[27] Beasley, J. E., “Route-first cluster-second methods for vehicle
routing. Omega 11(4),1983, pp. 403–408.
[28] Altinkemer, K. and B. Gavish, “Parallel savings based heuristics for
the delivery problem”. Operations Research 39(3), 1991,
[29] Ryan, D. M., C. Hjorring, and F., “Extensions of the petal method for
vehicle routing:, The Journal of the Operational Research Society
44(3), 1993, pp.289–296.
[30] Thompson, P. and H. Psaraftis, “Cyclic transfer algorithms for
multivehicle routing and scheduling problems”. Operations Research
41(5), 1993, pp. 935–946.
[31] Potvin, J.-Y. and J.-M. Rousseau, “An exchange heuristic for routing
problems with time windows. Journal of the Operational Research
Society 46(12), 1995, pp.1433–1446.
[32] Bramel, J. and D. Simchi-Levi, “A location based heuristic for
general routing problems”. Operations Research 43(4), 1995,
Engineering Letters, 18:2, EL_18_2_09
(Advance online publication: 13 May 2010)


[33] Renaud, J., F. F. Boctor, and G. Laporte, “A fast composite heuristic
for the symmetric traveling salesman problem”. INFORMS Journal
on Computing 8(2), 1996, pp.134–143.
[34] Kindervater, G. A. P. and M. W. P. Savelsbergh, “Vehicle routing:
handling edge exchanges”, 1997.
[35] Osman, I. H., “Metastrategy simulated annealing and tabu search
algorithms for the vehicle routing problem. Annals of Operations
Research 41(1-4), 1993, pp.421–451.
[36] Taillard, E. D., “Parallel iterative search methods for vehicle routing
problem. Networks 23(8), 1993, pp.661–673.
[37] Gendreau, M., A. Hertz, and G. Laporte, “A tabu search heuristic for
the vehicle routing problem”. Management Science 40(10),1994, pp.
[38] Van Breedam, A., “Improvement heuristics for the vehicle routing
problem based on simulated annealing”. European Journal of
Operational Research 86(3), 1995, pp. 480–490.
[39] Rochat, Y. and E. Taillard, “Probabilistic diversification and
intensification in local search for vehicle routing”. Journal of
Heuristics 1, 1995, pp.147–167.
[40] Xu, J. and J. Kelly , “A network flow-based tabu search heuristic for
the vehicle routing problem”. Transportation Science 30(4), 1996,
pp. 379–393.
[41] Kawamura, H., M. Yamamoto, T. Mitamura, K. Suzuki, and A.
Ohuchi, “Cooperative search based on pheromone communication for
vehicle routing problems”. IEICE transactions on fundamentals of
electronics, communications and computer sciences E81-A(6), 1998,
[42] Bullnheimer, B., R. Hartl, and C. Strauss, “An improved ant system
algorithm for the vehicle routing problem”. Annals of Operations
Research 89, 1999, pp.319–328.
[43] Toth, P. and D. Vigo (2003, Fall).2003, “The granular tabu search
and its application to the vehicle routing problem”. INFORMS
Journal on Computing 15(4), 2003, pp.333–346.
[44] Baker, B. M. and M. A. Ayechew, “A genetic algorithm for the
vehicle routing problem”. Computers & Operations Research 30(5),
2003, pp.787–800.
[45] Mazzeo, S. and I. Loiseau, “An ant colony algorithm for the
capacitated vehicle routing”. Electronic Notes in Discrete
Mathematics 18(1), 2004, pp.181–186.
[46] R.W. Floyd, “Algorithm 97 shortest path.” Commun. ACM 5., 1967.
[47] C.F. Daganzo, “The Distance Traveled to Visit N points with a
Maximum of C Stops per Vehicle: An Analytic Model and an
Application.” Transportation Science 18 (4),1981, pp. 331-350.
[48] S. Ree, B.S Yun, “A two-stage heuristic approach for the newspaper
delivery problem” , Computers &Industrial Engineering 30 (3), 1996,
pp. 501-509.
[49] A.P. Hurter, M. Van Buer, “The Newspaper Production/Distribution
Problem: Medium Sized Newspaper, The Journal of Business
Logistics, 17(1), 1996, pp. 85-108.
[50] M. Van Buer, D. Woodruff, R. Olson, “Solving The Medium
Newspaper Production/Distribution Problem.” European Journal of
Operational Research, 115, 1999, pp. 237-253.
[51] A.E. Carter, C.T. Ragsdale, “Scheduling Pre-printed newspaper
inserts using genetic algorithms.” Omega 30, 2002. pp.415-421.
[52] S.H. Song, K.S. Lee, G.S. Kim, “A Practical Approach to Solving a
Newspaper Logistics Problem Using a Digital Map. Computers &
Industrial Engineering 43 (1/2), 2002, pp.315-330.
[53] R. Russell, W. Chiang, D. Zepeda, “Integrating Multi-product
Production and Distribution in Newspaper Logistics” Computers &
Operations Research 35 (5), 2008, pp.1576-1588.
[54] R.J. Mantel, M. Fontein, “A Practical Solution to a Newspaper
Distribution Problem.” International Journal of Production
Economics 30/31, 1993, pp. 591-599.
[55] C. Kelnhofer-Feeley, Q. Hu, S. Chien, L. Spasovic, & Y. Wong, “A
Methodology for Evaluating of School Bus Routing: A Case Study of
Riverdales, New Jersey, Transportation Research Board 80
eeting, Woshingtone DC” Jauary 2001.
[56] S. Thangiah, B. Wilson, A. Pitluga, & W. Mennell, “ School Bus
Routing in Rural School Districts. Working Paper, Artificial
Intelligence and Robotics Lab. Computer Science Department ,
Slippery Rock University, Slippery Rock, PA.
[57] B.Adenso-Diaz, M. Gonzalez, and E. Garcia. ‘ A Hierarchical
Approach to Managing Dairy Routing” Interfaces, 1998, vol.28,
[58] N. Suthikarnnarunai, E. Olinick, “Improving Transportation Services
for the University of Thai Chamber of Commerce: A case study on
solving the mixed-Fleet Vehicle Problem with split Deliveries”
IAENG transaction on Engineering Technologies , Volume 1, 2008,
pp 200-211.
[59] L. Foulds, S. Wallace, J. Wilson, & L. Sogvolden, “Bookmobile
Routing and Scheduling in Buskerrud Country” Norway, Proceeding
of the 36
Annual Conference of the Operational Research Society of
ew Zealand, Christchurch, New Zealand, December 2001, pp.67-77
[60] A. Rodrigues, J. Ferreira, “Solving the Rural Postmand Problem by
Memeric Algorithms” Processing of MIC’ 2001 -4
nternational Conference, Porto, Potugal, July 2001.
[61] D. Burchett, E. Campion, “Mix Fleet Vehicle Routing Problem- An
Application of Tabu Search on The Grocery Delivery Industry.
Research Problem, University of Canterbury, New Zealand, 2002.
[62] BI. Kim, S. Kim, S. Sahoo, “ Waste Collection Vehicle Routing
Problem with Time Window”, Computer and Operational Research
33, 2006, pp.3624-3642.
[63] R.Dillmann, B.Becker and V. Beckefeld, “ Practical Aspect of Route
Planning for Magazines and Newspaper Wholesalers” European
Journal of Operational Research 90, 1996, pp 1-12.
[64] A. Harrison, and R. Van Hoek. “Logistics Management and
Strategy,” 3
ed. England:Prentice, 2008, pp. 141.
65] P. Toth, D. Vigo, “ The vehicle Routing Problem” Siam Monographs
on Discrete Mathematics and Application, 2002.
[66] C.K.Y. Lin, : Cooperative Strategy for a Vehicle Routing Problem
with Pick Up and Delivery Time Window”, Computer and Industrial
Engineer 55, 2008, pp. 766-682.
[67] A.L Guiffrida, R.Nagi, “Cost Characterizations of Supply Chain
Delivery Performance”. International Journal of Production
Economics, 102, 2006, pp.22-36.

Engineering Letters, 18:2, EL_18_2_09
(Advance online publication: 13 May 2010)